Keep your account alive.
Earn your payout.
You're one trade from a violation. Your firm's rules, as you trade, so you see the line before you cross it.
The stakes
<1%
of day traders turn a consistent profit
7%
of prop firm accounts ever reach a payout
~50%
of blown accounts die to the loss limit, not a losing strategy
The line kills more accounts than the strategy does. Cobalt shows you the line before you cross it.
Day-trader profitability: Barber, Lee, Liu & Odean (Taiwan, 1992-2006). Prop firm payout and failure rates: industry estimates, 2026.
7 day free trial
We email before the first charge
30 day refund on annual
No support tickets, no calls
Cancel any time
From settings, one click
Matches your broker
Reconciled to the cent on import
No AI credit meter
Recaps come with the plan
Supported prop firms
Apex 4.0 and TradeDay are live today. New firms ship rolling. Tell us yours.
01 · Protection
Your firm changes a rule. So does the engine.
Apex 4.0 and TradeDay, coded from the public rulebook. Every change published with a date and a link, and new firms ship rolling.
Live drawdown gauges
Your cushion to the floor, refreshed on every fill.
$2,900
Per firm rule engine
Each firm coded to its own rulebook.
Violation prediction
How many more losses before you cross the daily limit.
Payout readiness
Min days, profit threshold, consistency. The gates that release the payout, tracked live.
02 · Intelligence
AI that reads your behavior, not your tags.
Every claim cites a trade ID and clears a sample-size gate, nothing fires off three trades. And it reads your behavior against your firm's live rules, so the flag is “sizing up 60% into your daily-loss limit,” not generic FOMO.
- Daily recap of what you did and what to fix tomorrow.
- Weekly pattern surfacing across setup, time of day, instrument.
- Monthly behavioral summary tied to specific trade IDs.
- Anti hallucination validator rejects invalid citations.
A generic AI recap
- You traded emotionally today.
- Your win rate could be better.
- Consider managing your risk.
Cobalt, same week
You make money on first hour MES trades. You give it back after lunch.
Of 23 closed trades this week, your 9:30 to 10:30 ET trades returned +8.2R on a 61% win rate. After 1:00 ET your edge collapses: -4.1R on a 34% win rate, mostly chasing MNQ reversals you skipped in the morning.
This week: sit out after 1:00 ET. Cutting the afternoon turns last week from +4.1R net to +8.2R.
The receipts
The boring journal stuff, done right.
Manual entry, CSV import (Tradovate and NinjaTrader), screenshots, playbooks, tags, every analytics view. The difference is what reads it: a rule engine that knows your firm's rules and a coach that cites the exact trade.
Trifecta Score
83/100Based on 142 closed trades over the last 90 days
Focus on Profit factor78/100
Gross wins are 1.78× gross losses — solid edge, room to scale winners.
May 2026
+$8,330
19 trading days
15W · 4L
May 1
+$320
May 2
+$480
May 3
No trades
May 4
+$1,450
May 5
$180
May 6
+$220
May 7
+$540
May 8
$260
May 9
+$380
May 10
No trades
May 11
+$1,700
May 12
+$220
May 13
$440
May 14
+$380
May 15
+$240
May 16
No trades
May 17
No trades
May 18
+$900
May 19
+$520
May 20
$120
May 21
No trades
May 22
+$420
May 23
+$1,380
May 24
No trades
May 25
+$180
Am I improving?
Avg win, avg loss, and win rate over the last 8 weeks
Integrations
Tradovate and NinjaTrader via CSV today. More rolling out.
Tradovate
CSV today
NinjaTrader
CSV today
IBKR
Rolling out
Schwab
Rolling out
ThinkOrSwim
Rolling out
Reconciliation
Your journal should match your broker. To the cent.
The fastest way to lose a trader is a P&L that doesn’t match their statement. So every import is reconciled before it’s stored, and the math fails the build if it drifts.
Reconciled on every import
Every stored trade satisfies gross minus fees equals net, to within a cent. The invariant is enforced in CI on every import path, so a broken importer never ships.
Tested against real statements
Each broker format ships with a golden fixture: a real, anonymized CSV paired with the broker's own daily totals. If our math diverges by a cent, the build goes red.
You confirm it, every time
After an import we ask one question: does this match your broker? Your yes is the runtime check no amount of testing can replace.
Cobalt vs the alternatives
A trading journal logs the trade. Cobalt watches the line.
The category leader counts what each challenge cost you, after the fact. Cobalt sits on the planner and watches the line as you trade.
A generic journal
Built for every trader. Logs the trade after the fact.
Cobalt
The prop firm trader's journal. The protection lane.
- Your firm's rules, on the planner
Generic journal
Not built for thisCobalt
A live rule engine per firm, coded from the public rulebook. - Drawdown floor, as you trade
Generic journal
Not built for thisCobalt
The trailing floor on the planner, before you place the trade. - Daily loss limit
Generic journal
Logged after the session endsCobalt
A live countdown: how many more losses before you cross it. - Behavioral patterns
Generic journal
You self-tag emotions from a dropdownCobalt
Tilt and revenge, read from your data and scored against your live rule cushion. - Payout readiness
Generic journal
Spent vs earned, cost per funded accountCobalt
A gauge against the gates that release the payout: days, profit, consistency. - Onboarding
Generic journal
Built for every trader, tuned for noneCobalt
Pick your firm and stage; the rules load on day one.
Proof, not testimonials
Closed beta. So we show our work.
Built by traders who have hit the daily loss limit. Most journals plot your equity curve after the account is gone. Cobalt models your firm's rules as you trade, so you see the line before you cross it.
Ryan, founder of Cobalt
Reconciled to the cent
Gross minus fees equals net to the penny, or CI fails the build.
Insights cite the trade
Every AI claim names a trade ID and clears a sample-size gate.
Your data, isolated
Row-level security on every table, keyed to your account.
Pricing
Cheaper than the account it protects.
Pro $29/mo. Prop Firm $39/mo. Lose one funded account and you lose every payout it had left.
Free
Start journaling with the essentials.
$0
Pro
For traders running their own capital.
$290 billed yearly · 30-day money-back
- Unlimited trades
- Full trading journal: screenshots, playbooks, tags, notes
- Daily + weekly + monthly AI recap
- Full analytics: per-tag, per-emotion, R-distribution
- All CSV broker formats
- 7-day full trial
- 30-day money-back on annual
Prop Firm
RecommendedEverything in Pro, plus the prop firm command center.
$390 billed yearly · 30-day money-back
Keep one account alive and it's paid for itself.
- Everything in Pro
- Unlimited prop firm accounts
- Live rule engine: Apex 4.0 and TradeDay verified today. TopStep, FTMO, Tradeify, Leeloo, MyFundedFutures rolling out.
- Cushion gauges + daily-loss meter + payout readiness
- Pre-trade risk preview
- Public rule changelog access
Questions, answered
FAQ.
Keep your account alive
Cobalt is in closed beta. Get on the list.
We onboard traders in batches. Request an invite and we will reach out when your spot opens.